Annals of Occupational Hygiene Advance Access originally published online on January 13, 2005
Annals of Occupational Hygiene 2005 49(4):351-358; doi:10.1093/annhyg/meh095
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Original Article |
Comparison of Quartz Standards for X-ray Diffraction Analysis: HSE A9950 (Sikron F600) and NIST SRM 1878
Health and Safety Laboratory, Harpur Hill, Buxton, Derbyshire SK17 9JN, UK
* Tel: 01298 218555; fax: 01298 218571; e-mail: jim.chisholm{at}hsl.gov.uk
ABSTRACT
A further comparison of the Health and Safety Executive (HSE) standard quartz, A9950 (Sikron F600), and the National Institute of Standards and Technology (NIST) Standard Reference Material (SRM) 1878, standard respirable 
-quartz, has been carried out for the four principal diffraction peaks. In the earlier comparison by Jeyaratnam and Nagar (1993, Ann Occup Hyg; 37: 167–79), the standards were both treated in ways which might change the particle size distribution and therefore the proportion of crystalline quartz. The two standards have now been compared in the most direct way possible with the minimum of sample treatment. There are no significant differences in the diffraction peak positions for the two standards. Nor do the peak area intensities differ significantly. The peak height intensities are consistently and significantly higher for Sikron F600 than for NIST SRM 1878. The particle size broadening of the diffraction peaks is evidently greater for NIST 1878, whose mass median diameter is quoted as 1.6 µm against 2.6 µm for Sikron F600. Taking the certified reference value for SRM 1878 as 95.5 ± 1.1% crystalline quartz, the HSE standard A9950 (Sikron F600) contains 96.3 ± 1.4% crystalline quartz based on a comparison of peak area intensities. On the same basis but using peak height intensities, the nominal crystalline quartz content of A9950 (Sikron F600) is 101.2 ± 1.8%. Results obtained by comparison of quartz standards may not be generally applicable because of the effect of sample treatment on particle size and crystalline quartz content.
Keywords: crystalline silica • quartz • X-ray diffraction
INTRODUCTION
Inhalation of dust containing crystalline silica has long been recognized as a cause of pulmonary fibrosis (silicosis). Measurement of the airborne concentration of respirable crystalline silica (RCS) is required to provide the exposure data, which are essential for the successful implementation of exposure limits and control measures and for a proper assessment of the health risk.
The health risks from occupational exposure to RCS have recently been reviewed by the UK Health and Safety Executive (HSE, 2002
, 2003a) and the National Institute for Occupational Safety and Health (NIOSH, 2002). Data on the exposure–response relationship for silicosis suggest that the risk of developing silicosis post-exposure at the present British maximum exposure limit (0.3 mg/m3) is higher than had previously been thought (HSE, 2002, 2003b). NIOSH (2002) concluded that the results of its review supported the need for a revision of the Occupational Safety and Health Administration (OSHA) and Mine Safety and Health Administration (MSHA) exposure limits and continues to recommend an exposure limit of 0.05 mg/m3. HSE believes that most industries should be able to control exposure to within 0.1 mg/m3 (HSE, 2003b) and the British exposure limit is under review. Lowering of exposure limits implies measurement at lower airborne concentrations than at present and will impose more stringent demands on analytical methods. Reliable measurements are essential and depend on the availability of well-characterized reference standards. The National Institute of Standards and Technology (NIST) Standard Reference Material (SRM) 1878, standard respirable -quartz, is undoubtedly the best characterized standard, but others have been and are still being used. We need to know what the differences are among the various quartz standards if we are to relate airborne concentration measurements from different laboratories and make valid comparisons.
For many years the HSE has used a standard quartz, A9950 (Sikron F600), for analytical calibrations, testing of measurement and sampling methods, and quality control. This standard has been supplied to other laboratories for some time, but demand for it increased in the late 1990s, possibly because of poor availability at that time of the NIST SRM 1878. Many users of the HSE standard would like to have a certificate for that standard in order to meet quality assurance and accreditation requirements. A more rigorous comparison of the two standards was carried out to provide better data for such a certificate.
BACKGROUND
The HSE quartz standard identified as A9950 was originally prepared and characterized at the former Safety in Mines Research Establishment (SMRE) laboratory (Harris, 1984). After an initial assessment, 250 kg from a single batch of Sikron F600 was purchased from Quartzwerke, Frechen, Nordrhein-Westfalen, Germany. So far as is known, Sikron F600 is prepared by dry grinding with Silex and alumina pebbles and cyclone elutriation at the Frechen open pit.
A major factor in the choice of Sikron F600 was that its particle size and size distribution are very similar to those of the ashed respirable fractions of coal dusts from many British mines. Ideally a quartz standard should have a similar particle size and size distribution to the samples being analysed. Because it was originally intended for use in the analysis of dust recovered from filters, A9950 was chosen to have a size distribution which matched that of the dust after elutriation. It subsequently came to be used to prepare calibration filters by dispersal and sampling of the resulting aerosol as in MDHS (Methods for the Determination of Hazardous Substances) 37 (HSE, 1987) and MDHS 51/2 (HSE, 1988), for which a standard with a size distribution similar to a workplace dust (before elutriation) would have been more appropriate. However, no such standard is available and, if the dusts produced by different types of work have markedly different particle size distributions, a single standard may not always be suitable. So the use of the SMRE standard has become customary in UK laboratories even though it is not ideal for this purpose. The NIST SRMs 1878 and 1878a are no more suitable for the UK methods: they are described as ‘respirable’ and are finer than A9950 (see below). They are more suitable for NIOSH methods 7500, 7601, 7602 and 7603 (NIOSH, 1994), for all of which a respirable size standard is appropriate since no size selection takes place during calibration. Calibration as in MDHS 51/2 (HSE, 1988) will modify the size distribution of the already respirable standard, increasing the proportion of smaller particles and reducing the proportion of larger particles.
The Sikron F600 was supplied in 10 bags. Samples were taken from the top, middle and bottom of each bag. The homogeneity of the material was established and quantified by X-ray diffraction (XRD), infra-red spectroscopy and Coulter-counter measurements (Harris, 1984).
The net peak height intensity of the most intense quartz peak, 101 at 3.34 A, from each sample was measured. No indication of inhomogeneity was found (Harris, 1984). The coefficient of variation of the 30 samples was 1.9%; from previous work, variations in sample preparation were expected to give a coefficient of variation around 1% and other errors were expected from counting statistics and difficulties in locating the exact peak position.
The intensities of the infra-red absorption peaks at 800 and 780 cm–1 were also measured for the 30 samples and also gave no indication of inhomogeneity. The coefficients of variation were 3.3 and 2.3%, respectively, which were no more than expected for repeat determinations on the same material (Harris, 1984).
Coulter-counter measurements of the particle size distribution in terms of volume concentration were also carried out on the 30 samples. The size distributions of these samples showed good consistency (Harris, 1984).
Shortly after the NIST standard (SRM 1878) became available, Jeyaratnam and Nagar (1993) compared it with the HSE standard by looking at the differences which emerged when both were used for calibration for the methods used at the Occupational Medicine and Hygiene Laboratory (now the Health and Safety Laboratory, HSL) for analysis of quartz on filters and in bulk samples. Their work was intended mainly to consider the implications for routine analyses of using the new NIST standard, and from that standpoint it was satisfactory. However, despite its title, their article does not actually compare the two standards as such since they were modified during the sample preparation. For calibration for on-filter analysis, the standard was dispersed as an aerosol and sampled using a cyclone elutriator, which will inevitably modify the particle size distribution. For bulk analysis calibration, the standard was ground with nickel oxide as internal standard and 
-alumina as abrasive and diluent; the grinding will alter the mean particle size as well as the size distribution. So in each case, the actual comparison was of material derived from the standards by a specified treatment and not of the standards themselves. The quartz contents quoted by Jeyaratnam and Nagar (1993) for the HSE standard are valid when it is used for calibration by the two methods specified but not, strictly speaking, for other purposes.
The objective of the present work was to try to compare the two standards in the simplest and most direct way possible, avoiding any treatment which might modify them.
EXPERIMENTAL METHOD
Sample preparation
The method used was one originally developed at SMRE (Graham Revell, personal communication). The hollow of an XRD sample holder is slightly overfilled from the top with the powder to be examined. The surface of the powder is then levelled off by gently rubbing the roughened surface of a perspex block in a circular motion over the top surface of the sample holder without direct pressure. Any excess powder is wiped off the edges of the sample holder.
The validity of the method for comparing two materials depends on obtaining a reproducible packing of the powder. The reproducibility is established by making measurements on a series of diffraction mounts prepared in this way.
In fact, two types of sample holder were tried: (1) the standard Philips (thin) type for pressed bulk powders; (2) a type made at HSL intended for use with filters. A microscope cover slip was fixed in the filter position as a backing and the holder filled from the other side as described above.
A total of 20 diffraction mounts were made and measured for each standard, 10 for each type of sample holder.
XRD measurements
The measurement conditions used are summarized in Table 1.
|
For each quartz diffraction peak in turn, the background was calculated and subtracted using the default conditions of the DIFFRAC-AT software. Using the ‘Compute single peak’ option, the screen cursor was used to define the background points at the limits of the peak. The output then gives, inter alia, the peak position, peak height, peak area and full width at half-maximum (FWHM), which were recorded in each case.
The measurements were carried out over a sufficiently short period of time for the fall-off in intensity of the X-ray tube to be insignificant and were therefore not normalized to an external reference, as would be necessary for measurements extending over a long period of time.
RESULTS
Initially the mean and standard error of the peak position (2
), peak height, peak area and FWHM were calculated for each quartz peak in each set of 10 measurements for the two types of sample holder and the two quartz standards. There were no significant or consistent differences in peak position, peak height, peak area or FWHM between the two types of sample holder. This was the case for all four quartz peaks and for both quartz standards. Subsequently, therefore, the measurements using the two types of sample holder were treated as a single data set for each standard.
The results for the two quartz standards are compared in Tables 2 –4. The peak positions for the two are not significantly different; the standard error in peak position is about 0.02°2. However, the peak positions for SRM 1878 are consistently about 0.01°2 lower for all four quartz peaks than for those of A9950. This may indicate that the unit cell dimensions for A9950 are very slightly smaller than those for SRM 1878: if there were really no difference, the peak positions should be randomly higher and lower. An alternative explanation might be that the average sample displacement error for the two data sets is different.
|
|
|
The peak area intensities are a true measure of the amount of crystalline quartz present. The ratio of the peak areas HSE/NIST does not differ significantly among the four quartz peaks, nor does it differ significantly from 1.000. However, it is greater than 1.000 for all four peaks: if there were no difference between the two standards, the ratio should be randomly greater or less than 1.000.
The mean ratio of peak areas is 1.008 ± 0.009. The NIST SRM 1878 is certified as containing 95.5 ± 1.1% quartz and on that basis A9950 must contain 96.3% quartz. The standard error arising from the comparison is ± 0.9% and this is in addition to the error in the NIST certified figure. The combined error is ±1.4% and the quartz content of A9950 may be quoted as 96.3 ± 1.4%.
The results for peak height ratios are not the same: the ratio of the peak heights HSE/NIST does not differ significantly among the four quartz peaks but for three of them it does differ significantly from 1.000 (at the 5% level). For two of them, it differs significantly at the 1% level. It is also consistently greater than 1.000, the mean ratio being 1.060 ± 0.014, itself significantly greater than 1.000 at the 1% level.
Based on the NIST certified value (95.5 ± 1.1% quartz), the peak height ratios would correspond to a nominal quartz content for A9950 of 101.2% with an error of ±1.4% arising from the comparison and a combined error of ±1.8%. What this actually means is that A9950 has slightly sharper diffraction peaks and the lower integral breadth will lead to a higher peak height intensity for a given peak area.
For NIST SRM 1878, the FWHM is indeed higher for all four diffraction peaks (Table 3) but the difference is close to the borderline of significance (at the 5% level) for two of them. Since the peak height intensity can be estimated with greater precision than the FWHM, the peak height intensities should be a more sensitive indicator of differences in peak width than the FWHM.
DISCUSSION
The NIST standard consists almost entirely of particles in the respirable size range (<3 µm spherical equivalent diameter). That is not the case for A9950, which contains some coarser particles outside the respirable size range—about 20% by mass according to the size distribution data of Harris (1984). The difference in size distributions is shown in fig. 1 of Verma and Shaw (2001), in which the cumulative particle size distributions of various quartz standards are compared. It is well established (Demspter and Ritchie, 1952; Nagelschmidt et al., 1952; Gordon and Harris, 1955) that quartz particles have an amorphous surface layer, which makes up a higher proportion of the total material, the smaller the particles are. The apparently higher crystalline quartz content of A9950 may, therefore, be related to the reduced proportion of amorphous material when coarser particles are present.
In XRD, the integrated intensity (i.e. the peak area) is directly proportional to the amount of crystalline quartz present. Amorphous silica contributes only to the diffuse scattering, which, as part of the background, is subtracted to derive the peak area. As the exposure limits are for crystalline silica, it is that which must be determined and so a reference value for the percentage of crystalline quartz is essential for any standard. The NIST SRM 1878 standard is the only one for which such a value has been determined. In other standards that have been and are still being used, the amount of crystalline quartz, and therefore the peak area diffracted intensity, may vary depending on the source of the quartz and on any treatment that changes the proportion of amorphous silica present, that is, which changes the surface area (particle size) or otherwise alters the amount of amorphous material at the surface.
In infra-red spectroscopy, the absorbance is not so simply related to the amount of crystalline quartz present since it arises from transitions between energy states related to the vibration of Si–O bonds. Amorphous silica will contribute to the absorbance, but how much and in what way are not easily predictable. The absorbance may not be directly proportional to the amount of crystalline quartz present. Accordingly, Verma and Shaw (2001) in their comparison of quartz standards by infra-red spectroscopy are careful not to interpret their results in terms of the percentage crystalline quartz in each standard. Strictly speaking, it is invalid to use a percentage crystalline quartz derived from XRD comparison in infra-red spectroscopic analysis and this includes the reference values for SRM 1878, which were derived by Rietveld analysis of XRD data.
Peak intensity comparisons in relation to on-filter analysis
The method for on-filter analysis (Pickard et al., 1985) referred to by Jeyaratnam and Nagar (1993) uses peak area intensities. The results obtained here are probably consistent with their conclusion that the two standards do not differ significantly when used for calibration for on-filter analysis. In the present comparison, the crystalline quartz contents of the two standards do not differ significantly, although the crystalline quartz content is slightly higher for A9950. The sampling process used to prepare calibration filters will exclude the coarser particles of A9950 and make the size distributions of A9950 and NIST 1878 as deposited on the filter more alike. As the amorphous surface layer makes up a higher proportion of the smaller quartz particles, the size selection will lower the proportion of crystalline quartz in A9950 by slightly more than in NIST 1878 and so bring the two standards even closer together.
Kauffer et al. (2002) carried out a more rigorous comparison of some -quartz samples with NIST SRM 1878a, including a sample of Sikron F600 (referred to as QUIN2) also from Frechen but extracted many years later than the HSE material. It should be noted that the comparison in the present study, like that of Jeyaratnam and Nagar (1993), is with SRM 1878, the first batch of the NIST standard, certified as containing 95.5 ± 1.1% crystalline quartz, whereas that of Kauffer et al. (2002) is with SRM 1878a, the second batch, certified as containing 100.00 ± 0.21% crystalline quartz. As a consequence, the differences between Kauffer et al. (2002) and Jeyaratnam and Nagar (1993), although still significant, are not so great as might appear from a direct comparison of the two. The certified values for crystalline quartz content have been used to recalculate the ratios of peak areas relative to SRM 1878a (Table 4), which are then directly comparable with the ratios Crq from Kauffer et al. (2002).
The comparison by Kauffer et al. (2002) paid particular attention to the dependence of diffracted intensity on particle size distribution. In a similar way to Jeyaratnam and Nagar (1993), they prepared samples by dispersing the quartz in air and collecting dust on the filter using a cyclone sampler. However, by varying the flow rate, they were able to alter the size distribution of the sampled quartz particles, which they determined using a Grimm dust monitor, and so were able to study the variation of diffracted intensity per unit mass with particle size (volume median diameter). Their results are consistent with an interpretation in which the diffracted intensity per unit mass increases linearly with increasing volume median diameter, this increase being similar for all the quartz standards they examined. In order to relate their comparisons to the present one, it should be understood that their parameter, Crq, is the diffracted intensity per unit mass relative to SRM 1878a for any given volume median diameter, that is, along any vertical line on their plots of intensity against diameter (Kauffer et al., 2002, figs 5–8). So their interpretation is not a direct comparison of the standards but is equivalent to a comparison of material derived from each standard by size selection to the same volume median diameter. Otherwise their ratios Crq correspond to the recalculated peak area ratios in Table 4. The marked differences are to be expected given that the former compare size-selected fractions with the same volume median diameter and the latter compare the untreated standards.
Biggins (1982) determined the size distribution of HSE standard Sikron F600, the mass median diameter being 2.6 µm (weight average diameter 2.8 µm). A median particle size of 1.6 µm is quoted on the Certificate of Analysis for SRM 1878a; the certificate for SRM 1878 quotes the mass mean equivalent spherical diameter (1.62 µm). Although a direct comparison is not possible, it seems likely that the particle size distributions of SRM 1878 and 1878a are not markedly different. For a homogeneous material, the mass median and volume median diameters should be equivalent. The comparison with Kauffer et al. (2002) in Table 4 is therefore between Sikron F600 (HSE A9950) with volume median diameter 2.6 µm and SRM 1878a with volume median diameter 1.6 µm. Sikron F600 will have a higher diffracted intensity per unit mass because of its particle size, apart from any difference in amorphous silica content, and qualitatively this explains the higher ratios to SRM 1878a in the present comparison.
Quantitative comparison is more difficult: inspection of figs 5–8 in Kauffer et al. (2002) suggests that a 1 µm increase in volume median diameter will lead to an increase of about 2–3% in intensity per unit mass, which is not sufficient to account for the whole of the differences in Table 4.
The comparison is further complicated by the effect of agglomeration of particles during dust generation on the size distribution. Kauffer et al. (2002) note that the volume median diameters measured after cyclone elutriation were higher than those for the bulk powder, which implies agglomeration. The Certificate of Analysis for SRM 1878 notes that it agglomerates very easily during handling and storage, but this is not noted on the certificate for SRM 1878a. However, the effect of agglomeration is not simply what would be expected from the increase in particle size: agglomeration cannot change the proportion of amorphous material present and therefore should not change the diffracted intensity per unit mass. If the variation of diffracted intensity with particle size is caused by an amorphous layer, the intensity should depend on the volume median diameter before agglomeration, not after.
Much of the difference between the present comparison and that of Kauffer et al. (2002) can be explained by size differences between the standards and the difference between SRM 1878 and 1878a. But there does appear to be a small remaining difference of at least 3% for which no definite explanation can be given.
Some of the differences between the comparisons using infra-red spectroscopy carried out by Kauffer et al. (2002) and by Verma and Shaw (2001) can be accounted for by the difference in crystalline quartz content of SRM 1878 and 1878a and the difference in particle size between the SRM standards and Sikron F600 (HSE A9950). But again a residual difference remains unexplained.
Peak intensity comparisons in relation to analysis of bulk samples
The method for analysis of bulk samples (Taylor, 1978) referred to by Jeyaratnam and Nagar (1993) uses peak height intensities. Their conclusion that A9950 appears to contain 3% less quartz than SRM 1878 when used for this analysis is not at first sight reconcilable with the conclusion in the present study that the peak heights of A9950 are consistently and significantly higher by 5–8% than those of SRM 1878. Jeyaratnam and Nagar themselves appear to have been puzzled by the apparent difference in quartz content of A9950 revealed between bulk and on-filter analysis. They suggest in their conclusions that the explanation may lie in differences in the crystallinity of the two standards in the respirable size range and implicitly in the size-selective sampling in preparing calibration filters. In preparing calibration samples for bulk analysis, the standard quartz is ground for 30 min under cyclohexane (Taylor, 1978) with nickel oxide internal standard and -alumina abrasive. Much of the coarse material in A9950 will be reduced in size by the grinding and this will increase the surface area and the proportion of amorphous material, reducing the diffracted intensity. Particle size broadening of the diffraction peaks will also be greater for smaller particles and would reduce the peak height intensity. These effects may at least partially explain the difference in results; a comparison of the two quartz standards for analysis of bulk samples using both peak heights and peak areas might be instructive. Bhaskar et al. (1994) report significant differences in XRD peak height intensity between size fractions with different median diameters; Verma et al. (2002) also report clear differences in infra-red absorbance for unsieved material, <38 µm and <10 µm sieved fractions.
Peak widths
The greater peak widths observed for SRM 1878 are consistent with its generally lower particle size noted above. However, in the context of peak broadening, particle size refers to the size of the diffracting crystallites (Altree-Williams and Clapp, 2002), of which there may be many in each particle of a powder. There is, however, a crude correlation between particle size and crystallite size if only because the crystallites can never be larger than the particles.
Contributions to peak broadening may arise from two other sources: microstrain, which depends on deformation history and mechanical treatment, and disorder in the regular atomic arrangement which characterizes a crystal. There may be some loss of structural regularity in the crystal; for instance, there may be a gradation from a regular atomic arrangement into the more random one of an amorphous surface layer. Kauffer et al. (2002) suggest that some of the differences in infra-red absorbance at particular wave numbers for their samples may be related to global deterioration inside the quartz grains rather than separation into amorphous and crystalline phases.
Preferred orientation effects
The peak area ratios Sikron F600/SRM 1878a, Crq, found by Kauffer et al. (2002) are significantly higher for the 100 and 101 quartz peaks (at 20.8 and 26.6°2) than for the 112 and 211 peaks (at 50.1 and 59.9°2). The difference is consistent with the preferred orientation of quartz crystals with hexagonal prism faces of the form {10
0} and rhombohedral faces of the form {101} parallel to the filter, these being the commonest and largest growth faces of natural quartz. No such indications of preferred orientation have been found in the present comparison.
CONCLUSIONS
The unit cell dimensions of A9950 appear to be very slightly smaller than those of SRM 1878 but the difference in peak positions is not statistically significant.
Taking the certified reference value for SRM 1878 as 95.5 ± 1.1% crystalline quartz, the HSE standard A9950 (Sikron F600) contains 96.3 ± 1.4% quartz based on a comparison of peak area intensities. The difference in crystalline quartz content is not statistically significant but is consistent for all four quartz peaks measured. On the same basis but using peak height intensities, the nominal quartz content of A9950 (Sikron F600) is 101.2 ± 1.8%. The HSE standard A9950 (Sikron F600) has slightly sharper diffraction peaks than SRM 1878 and this leads to significantly higher peak height intensities (5–8% higher) for all four quartz peaks measured.
There are indications that the diffracted intensity obtained from a quartz standard may depend on the treatment it has been given, particularly if that treatment changes the particle size distribution and, as a consequence, the proportion of amorphous material present as a surface layer. Much of the difference between the present comparison and that carried out by Kauffer et al. (2002) can be explained in this way and by the difference between SRM 1878 and 1878a, but a small discrepancy may still remain. Results obtained by comparison of quartz standards may not be generally applicable because of the effect of sample treatment on particle size and crystalline quartz content.
The effect of sample treatment has significant implications for the use of secondary standards in quartz analysis since the reference value of the percentage crystalline quartz is based on a comparison with the primary standard. That reference value for the secondary standard is only valid if the sample treatment in the analysis exactly matches (in its effect on the percentage of crystalline material present) the treatment used in the comparison of the two standards. So a reference value for the secondary standard, which is based on one particular comparison, is not generally applicable and, for diffraction peak areas, may differ for different types of treatment, such as grinding, which change the proportion of crystalline quartz present. For peak heights, the reference value will in addition be sensitive to treatments that alter crystallite size and/or microstrain since these determine peak width.
The comparisons of Jeyaratnam and Nagar (1993) remain valid for the specific analytical methods and sample treatments used by them. Their results give an indication of how much difference sample treatment may make to the reference value for the secondary standard. For determination of quartz on filters using MDHS 51/2 (HSE, 1988), in which the sample is dispersed as an aerosol and collected on filters using cyclone samplers, Jeyaratnam and Nagar (1993) found no significant difference between the two standards. But for the analysis of bulk samples (Taylor, 1978), in which the sample is ground in a ball mill for 30 min, the HSE standard appeared to contain 92.5% crystalline quartz based on the reference value of 95.5% for NIST 1878, a difference of about 3%. However, the method uses peak height intensities for which 101.2% crystalline quartz is the nominal reference value found in the present comparison, implying a difference of about 9%. A variation similar to this but for peak areas is shown by figs 5–8 of Kauffer et al. (2002), who deliberately set out to modify the particle size by changing elutriation conditions. For all three standards examined by them, the diffracted intensity per unit mass varied by approximately ±8% from the mid-range value as the volume median diameter ranged from about 2.5 to 4.5 µm.
There are, of course, other sources of uncertainty in quartz analysis. Data from a review of the Workplace Analysis Scheme for Proficiency (WASP) results (Stacey et al., 2003) show how large the uncertainty is in quartz-on-filter analysis. Nearly all the laboratories in this scheme that analyse quartz on filters are in the UK, and most if not all of these use the HSE quartz standard, so differences between standards should contribute little or nothing to the variability. Excluding results from laboratories using indirect methods (i.e. recovering the dust from the filter), which showed the greatest variation, the overall relative standard deviation (RSD) was 8.5%. For laboratories using infrared spectroscopy and XRD, the within-laboratory repeatability, expressed as RSD, was 5.6 and 6.7%, respectively, for all samples and ranged from about 3 to 8% and 6 to 8%, respectively, depending on filter loading. The inter-laboratory reproducibility, also expressed as RSD, ranged from 5 to 7% for infra-red spectroscopy and from 7 to 10% for XRD. So at about 3–9%, the effect of different sample treatment is similar in magnitude to the random analytical error in quartz-on-filter analysis and is by no means small enough to be disregarded. More importantly, changes in crystalline content resulting from sample treatment will introduce a bias into any final result unless the reference value is valid for the analytical method used.
ACKNOWLEDGEMENTS
I am grateful to my colleague Graham Revell for invaluable assistance with information on the history of the SMRE standard quartz A9950, which later became the HSE standard, and for clarifying many points relating to particle size and elutriation.
Received April 19, 2004; in final form October 6, 2004
REFERENCES
Altree-Williams S, Clapp R. (2002) Specific toxicity and crystallinity of -quartz in respirable dust samples. AIHA J; 63: 348–53.
Bhaskar R, Li J, Xu L. (1994) A comparative study of particle size dependency of IR and XRD methods for quartz analysis. Am Ind Hyg Assoc J; 55: 605–9.[Web of Science][Medline]
Biggins PDE. (1982) Standard quartz samples compared by X-ray powder diffractometry and infra-red analysis, BCIRA report 1472. BCIRA J; 30: 235–53.
Dempster PB, Ritchie PD. (1952) Surface of finely grained silica. Nature; 169: 538–9.
Gordon RL, Harris GW. (1955) Effect of particle size on the quantitative determination of quartz by X-ray diffraction. Nature; 175: 1135–7.
Harris GW. (1984) A European quartz standard material for infra-red spectroscopy and X-ray diffraction analysis. Unpublished draft HSE paper.
HSE. (1987) Quartz in respirable airborne dusts. Laboratory method using infrared spectroscopy (direct method). Methods for the Determination of Hazardous Substances MDHS 37. London: Health and Safety Executive. ISBN 0 7176 0208 7.
HSE. (1988) Quartz in respirable airborne dusts. Laboratory method using X-ray diffraction (direct method). Methods for the Determination of Hazardous Substances MDHS 51/2. London: Health and Safety Executive. ISBN 0 11 883641 3.
HSE. (2002) Hazard assessment document. Respirable crystalline silica—Phase 1. Variability in fibrogenic potency and exposure–response relationships for silicosis. EH75/4. London: HSE Books. ISBN 0 7176 2374 2.
HSE. (2003a) Hazard assessment document. Respirable crystalline silica—Phase 2. Carcinogencicty. EH75/5. London: HSE Books. ISBN 0 7176 2191 X.
HSE. (2003b) Chemical Hazard Alert Notice 35—Respirable crystalline silica. Available from http://www.hse.gov.uk/pubns/chan35.htm (accessed 30 September 2004).
Jeyaratnam M, Nagar N. (1993) Comparison of -quartz standard Sikron F600 (HSE A9950) with the NIST respirable -quartz standard SRM 1878 for bulk and on-filter analysis. Ann Occup Hyg; 37: 167–79.
Kauffer E, Moulut JC, Masson A et al. (2002) Comparison by X-ray diffraction and infrared spectroscopy of two samples of quartz with the NIST SRM 1878a quartz. Ann Occup Hyg; 46: 409–21.
Nagelschmidt G, Gordon RL, Griffin OG. (1952) Surface of finely-grained silica. Nature; 169: 539–40.[CrossRef][Medline]
NIOSH. (1994) NIOSH manual of analytical methods (NMAM), 4th edn. DHHS (National Institute for Occupational Safety and Health) Publication 94-113. Available from http://www.cdc.gov/niosh/nmam/
NIOSH. (2002) NIOSH hazard review. Health effects of occupational exposure to respirable crystalline silica. DHHS (National Institute for Occupational Safety and Health) Publication 2002129.
Pickard KJ, Walker RF, West NG. (1985) A comparison of X-ray diffraction and infra-red spectrophotometric methods for the analysis of -quartz in airborne dusts. Ann Occup Hyg; 29: 149–67.
Stacey P, Tylee B, Bard D et al. (2003) The performance of laboratories analysing -quartz in the Workplace Analysis Scheme for Proficiency. Ann Occup Hyg; 47: 269–77.
Taylor M. (1978) Methods for the quantitative determination of asbestos and quartz in bulk samples using X-ray diffraction. Analyst; 103: 1009–20.
Verma DK, Shaw DS. (2001) A comparison of international silica (-quartz) calibrations standards by Fourier transform-infrared spectroscopy. Ann Occup Hyg; 45: 429–35.
Verma DK, Johnson DM, Des Tombe K. (2002) A method for determining crystalline silica in bulk samples by Fourier transform infrared spectrophotometry. Ann Occup Hyg; 46: 609–15.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  P. Stacey, E. Kauffer, J.-C. Moulut, C. Dion, M. Beauparlant, P. Fernandez, R. Key-Schwartz, B. Friede, and D. Wake An International Comparison of the Crystallinity of Calibration Materials for the Analysis of Respirable {alpha}-Quartz Using X-Ray Diffraction and a Comparison with Results from the Infrared KBr Disc Method Ann. Hyg., August 1, 2009; 53(6): 639 - 649. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. K. VERMA and D. S. SHAW Comparison of Crystalline Silica ({alpha}-Quartz) Calibration Standards NIST-SRM 1878 and NIST-SRM 1878a by Fourier Transform Infrared Spectrophotometry Ann. Hyg., June 1, 2005; 49(4): 359 - 361. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||